The Overdetermination Argument Revisited AGUSTÍN VICENTE Dpto. de Filosofía, Lógica y Filosofía de la Ciencia, Facultad de Filosofía y Letras, Universidad de Valladolid, Pl. del Campus, s/n. 47011 Valladolid, Spain; E-mail:
[email protected] Abstract. In this paper I discuss a famous argument for physicalism – which some authors indeed regard as the only argument for it – the “overdetermination argument”. In fact it is an argument that does not establish that all the entities in the world are physical, but that all those events that enter into causal transactions with the physical world are physical. As mental events seem to cause changes in the physical world, the mind is one of those things that fall within the scope of the argument. Here I analyze one response to the overdetermination argument that has acquired some popularity lately, and which consists in saying that what mental events cause are not physical effects. I try to show that recent attempts to develop this response are not successful, but that there may be a coherent way of doing so. I also try to show that there seems to be a philosophical “niche” in which this way might fit. Key words: dual explanandum strategy, mental causation, overdetermination argument, physicalism
1. Introduction In this paper I will discuss a famous argument for physicalism, and a particularly interesting response to it. Some authors (see, e.g. Peacocke, 1979; Papineau, 1990, 2000; Antony and Levine, 1997; Sturgeon, 1998; Levine, 2001) hold that the argument I am about to present is the argument for physicalism. Their basic reason for this claim of exclusivity is that the other main thread one may pull in order to argue for physicalism, namely, the one related to unificatory ideals, has been shown to be under threat. It used to be argued that physicalism was true because we had inductive reasons to think that all the sciences were becoming one, or, better, because for something to be explainable, it had to be capable of being integrated within a basic framework, i.e. because if the world was going to be intelligible, then it had to be unifiable. However, the reflection started by Fodor (1974) about the special sciences has arguably shown that there is no prospect for all the sciences to become one, or for physics to provide reducing explanations of the events belonging to other sciences, since each science proceeds at its own level of generalization and has its own taxonomic and explanatory interests. Thus, there seems to be no epistemological argument for physicalism. If you want to argue for physicalism you have to construe a straight ontological argument which shows that, no matter how the different sciences individuate and generalize, they are always ultimately speaking about physical properties. The argument to be discussed is quite straightforward, though it gets more sophisticated as one tries to find possible refutations of it. It is not supposed to Minds and Machines 14: 331–347, 2004. © 2004 Kluwer Academic Publishers. Printed in the Netherlands.
332
AGUST´IN VICENTE
establish physicalism tout court, i.e. the thesis that our world — and those worlds that are physically identical to ours — contains just physical entities, but physicalism of a restricted kind, namely, the thesis that all the entities that enter into causal interactions with the physical world are physical. The argument, thus, is of special concern to the mind, since at least some mental events (i.e. the instantiation of a mental property by an individual) are (intuitively) clearly involved in causal interactions with the physical world. In fact, the mind has classically been the focus of the present argument, and my discussion will also be concerned with it. That is, I will discuss whether the argument does or does not establish that mental events are physical. Before presenting it, however, I want to present what I take to be a reasonable doubt as to its status of uniqueness. The doubt arises from the fact that the most promising and interesting naturalist/physicalist accounts of the mind, such as Millikan’s and Dretske’s, construe mental properties as teleological. However, teleological properties do not have, as such, causal powers. But, of course, if the only reason for being a physicalist about the mind were that mental events cause physical events, there would be no point in becoming a naturalist/physicalist1 if you end up claiming that mental events cannot be causes, which is what Millikan (1993) proposes: “On the account of this essay, the semantic category of a thought is determined relative to its biological functions, which depend in turn upon its history, upon its place relative to certain prior events. But having a certain history is not, of course, an attribute that has “causal powers”. And again: “[...] that a thing has a teleofunction is a causally impotent fact about it. Especially, it is never directly because a thing has a certain function that it performs that function or any other function. More nearly the reverse is true. The thing exists and has a certain function because things homologous to it have performed that function (or better, that effect) in the past” (p. 136, her italics). The argument, well known now, is sometimes called ‘the exclusion argument’ but more often ‘the overdetermination argument’. It consists basically of three premises, namely: (i) the principle of the causal closure of the physical: every physical effect has a sufficient causal antecedent that is also physical; (ii) the causal efficacy of the mental: mental events produce changes in the physical world; (iii) no overdetermination: there is no mental-physical causal overdetermination. And the conclusion is that (iv) mental events are physical events. Now, this physicalist conclusion is not satisfactory for at least two reasons. According to the view on events assumed here, to say that mental events are physical events means that mental properties are physical properties2 . However, as is well known, mental properties cannot be identical to physical properties. First, mental properties are multiply realizable. Second, they have some very intriguing features
THE OVERDETERMINATION ARGUMENT REVISITED
333
that no physical property has. Some mental properties have “aboutness”, i.e. they have a representational content, and others are “qualitative”, i.e. they feel in some way or other. Suppose then that one is understandably unsatisfied with the physicalist conclusion provided by the overdetermination argument. What can be done? All escapes have been tried. Some philosophers have denied (i), i.e. that the physical world is causally closed. Here we find emergentists and some dualists (interactionist dualists). Another group of philosophers has defended that mental events are not causally efficacious, thus denying (ii) and arguing for epiphenomenalism. Finally, there are those who have argued that there is indeed mental-physical overdetermination, and have tried to block the argument at (iii). None of these “solutions” has been regarded as satisfactory by most of the philosophical community. First, the denial of the first premise, it is usually said, commits one to the denial of what seems to be a principle of contemporary physics. There is some room for dissidence here, though, because it is not entirely certain that physics supports this principle of causal closure. At least, it is not a law that appears in physics textbooks. However, one may argue that it is not a physical law, but rather a methodological norm that guides physicists in their research and is well supported by inductive evidence. Alternatively, it may be said that it is a law of physics, in the sense that it supervenes on, or is supported by, actual laws of physics, such as conservation laws. Conservation laws at least guarantee that every effect that involves an increase or decrease in a conserved quantity requires a corresponding decrease or increase in its cause. That is, perhaps from conservation laws you do not get the causal closure principle, but you get something that is enough for the overdetermination argument to work, for the effects that mental causes are said to bring about do involve increases and decreases in energy and quantity of movement. That is, contemporary physics runs counter to emergentism and interactionist dualism (minimally) via conservation laws (see on this point Papineau, 2000). Second, denying that mental events are causes – that is, becoming an epiphenomenalist – is both counterintuitive and dangerous. It is counterintuitive because we do experience the causal efficacy of our thoughts, desires and (also) qualia states all the time, both as causes of behaviors when we move our body in accordance with our desires, and as causes of other mental states, as when we go through a modus ponens. It is dangerous because a sensible criterion of existence has it that “to be is to have causal powers” (Alexander’s dictum, made popular by Kim, 1993, and used also by Armstrong, 1997). Thus, epiphenomenalism is a sure road to eliminativism, and the denial of the causal efficacy of the mental turns out to be useless as a response to the exclusion argument for physicalism. Finally, claiming that the movements of our bodies are causally overdetermined by physics and the mind (denial of (iii)) commits one to the existence of unexplained and massive coincidences. Or, if you want to remove the mystery of these coincidences, and say that the constant overdetermination is due to the presence of
334
AGUST´IN VICENTE
psychophysical laws (as Mellor and Crane, 1990, do), then the commitment is to the existence of unexplained and unnecessary laws. Now, suppose the three “classical” escapes from the overdetermination argument are as impracticable as I have suggested, and that physicalism is still unwelcome: is there any way to block the argument that we have not yet considered? Let’s get back to premise (ii). We have considered one way of denying it, namely, by saying that mental events are epiphenomenal. However, it can be rejected on other grounds. Instead of saying that mental events are inert, it is open to the critic of physicalism to claim that mental events are causally potent, only that they do not cause events that are physical, or at least that they are such that the causal closure principle does not “rule” over them. This is the “dual explanandum strategy”. Two critics of physicalism have lately made use of this strategy: Scott Sturgeon (1998, 1999) and Jennifer Hornsby (1997)3 . What I will do in what follows is present and criticize their respective accounts. I will try to show that the strategy is highly demanding, more so than their proponents seem to be aware. In order to do this, I will revisit first Sturgeon’s and then Hornsby’s proposals. After that, we will reach what I take to be an adequate account of what the “dual explanandum strategy” must consist in. As I say, this account exacts a high price, and as such this antiphysicalist response may be easily seen as counterintuitive. However, I will suggest that the picture that finally emerges as a response to the demands of the strategy may find an echo in some neo-existentialist positions coming from the field of robotics.
2. The Dual Explanandum Strategy 2.1. THE OVERDETERMINATION PROBLEM AGAIN I will start by posing a problem to the defenders of the dual explanandum strategy so that it begins to be clear what such a defender must be committed to. The problem is none other than the overdetermination argument again, but now a stronger version of premise (iii) is used, namely: (iii ) there is no mental-physical (causal or of any other kind) overdetermination. Kim (1998) makes use of a premise like this in step (vi) in his argument from supervenience, destined to force downward causation. There he is considering mental-to-mental causation, and he asks: how can the instantiation of a mental property cause the instantiation of another taking into account that the instantiation of the second mental property can be explained by the instantiation of a physical property that subvenes it? Answer: only by causing such a physical property. The case I want to consider is no different. For suppose I say that mental events cause behavioral events, which are not identical to any kind of physical events. However, as it happens, every time a behavioral property is instantiated, a given physical property gets instantiated simultaneously, and the instantiation of this physical
THE OVERDETERMINATION ARGUMENT REVISITED
335
property is sufficient for the instantiation of the behavioral property. If this is the case, then we are forced to say that mental events cause behavioral events only by causing some physical events, and the strategy is thereby dismantled. Thus, it seems crucial for the defender of the dual explanandum approach to maintain that behavioral events do not supervene on physical properties.
2.2. THE MACRO AND THE MICRO AND THE CAUSE AND ESSENCE PRINCIPLE Scott Sturgeon (1998, 1999) has defended a version of the dual explanandum strategy that accords to the constraint I have just presented. His view is first that macro-physics is not causally closed, second, that behavioral events belong to the macro-image and third, that micro events are, as far as we know, not sufficient for macro events. I will explain these in turn. First of all, Sturgeon distinguishes between quantum-mechanical completeness and broadly physical completeness. This distinction gives us two different arguments, depending, that is, on whether we have (i ) every quantum effect has a sufficient causal antecedent that is also quantummechanical or (i ) every broadly physical effect has a sufficient causal antecedent that is also broadly physical. Now the problem for the physicalist, according to Sturgeon, is that while (i’) is true, nobody holds that mental events have quantum effects. Rather, what both science and common sense claim is that mental events have broadly physical effects. But, for that same reason, (i”) seems to be false. In other words, contemporary physics only backs (i’), which does not affect mental causation, since mental events cause just broadly physical events. However, things are not so easy for the antiphysicalist. The defender of the argument from exclusion, Sturgeon admits, can invoke the fact that quantum events compose macro events in order to impose a completeness principle. This can be obtained by holding (p. 417) (a) Closure under downward-composition: If C causes E and E is composed by E∗, then C causes E∗. (b) Closure under upward-composition: If C causes E and E composes into E∗, then C causes E∗. Then, as every macro event is composed by quantum events and every quantum event has a complete quantum cause, then every macro event has a complete quantum cause (by (b)). Also, as every macro event is composed by quantum events, then (by (a)) some macro causes cause quantum events (and thus compete with quantum causes). But this is not the end of the story. Principles (a) and (b) are false as they stand. Sturgeon presents three counterexamples to them of increasing strength, and moves on to offer a rewriting. The basic idea now is that causal relations can move across
336
AGUST´IN VICENTE
different levels of reality only if there is a relation of essence inclusion between the effects (see Yablo, 1992). Then, principles (a) and (b) can be true when, and only when, the relation of composition is a relation of essence inclusion. In Sturgeon’s words (p. 422): “The moral is this: intuition sees causation flowing across composition to just the degree it sees composition respecting an effect’s essence. Intuition reflects our commitment to the Cause-and-Essence principle: (C&E) C causes E iff C is sufficient to bring about what is essential to E.” Stemming from this principle, we have: (a ) If C causes E and E is essentially composed by E∗, then C causes E∗. (b ) If C causes E∗, and E∗ essentially composes into Ethen C causes E. So the reasoning must be the following: if every macro event is essentially composed by quantum events and every quantum event has a complete quantum cause, then every macro event has a complete quantum cause. Also, if every macro event is essentially composed by quantum events, then (by (a)) some macro causes cause quantum events (and thus compete with quantum causes). The role of principle (a ) is, as can be seen, similar to that played by the “generalized no overdetermination” premise in the last section. There, what the physicalist wanted to obtain was the commitment to downward causation, and this can be done by pressing the defender of the dual explanandum strategy either with the denial of overdetermination (iii ) or with principle (a ). The difference is that this second principle gives rise to an argument that is considerably tamer: for the argument from (iii ) to work we only need that parts E∗ be sufficient for E, whereas here we require E to be a mereological whole made of essential parts. That’s a far stronger condition, which generates an argument with a much narrower range of application. It is also interesting to contrast principle (b ) with a hypothetical argument for upward causation starting from the denial of any kind of overdetermination. There are fewer parallelisms here, as can be noted. First, the argument using the generalized no overdetermination premise (iii ) cannot provide an equivalent to (b ). For that we would need the instantiation of E to be sufficient for the instantiation E∗. Under that assumption it would be possible to say that since E∗ cannot have two sufficient conditions, E and C, then C must have caused E. But even if E∗ essentially composes into E, that does not mean that E is sufficient for E∗: to use Sturgeon’s example, certain duck movements essentially compose a duck flock movement, but the presence of that duck flock movement is not sufficient for the presence of those duck movements. On the other hand, and much more important, the argument from the generalized no overdetermination premise commits the physicalist to something she would never accept, namely, that micro events cause other micro events via the instantiation of macro properties. But of course, all this is good news for the physicalist: her position is now stronger, for she has a good argument for downward causation (which she previously had) and another for upward causation (which she did not).
THE OVERDETERMINATION ARGUMENT REVISITED
337
Now let’s see how Sturgeon tries to resist the argument he has construed. Of course, he admits the Cause-and-Essence principle, and its corollaries (a ) and (b ). What he rejects is that the composition of quantum events are essential for macro events. And this he justifies by pointing to the conceptual gap that there is between the macro and the micro images. Sturgeon’s denial of the supposition that the quantum world composes essentially the macro world applies not only to relations between objects but also to relations between spatial events. The argument is that macro events are conceptually remote from quantum events: “The conceptual divide between quantum and macro reality is great. We should be cautious in accepting unrestricted causal flow between the two. For we cannot see how quantum events build into the splendor of macro reality” (p. 424) In the particular case of spatial events (position and movement), this conceptual divide comes to the following. Quantum spatial events have two intriguing characteristics, superposition and projection. According to the superposition effect, a particle’s position when it is not being measured is best conceived as a wave-like phenomenon that has a certain probability of being at various positions. According to the measurement effect, on the other hand, when measured the wave collapses into a particle in one particular position. Needless to say, none of these striking effects govern macro spatial events. What is more, we cannot understand how things that accord to them can give rise to the determinacy we observe in the macro world. So this is the argument for the conceptual gap. As said above, Sturgeon goes from here to the claim that there is (perhaps) no essence inclusion relationship, hence no causal relations, between quantum events and macro events. The idea here seems to be that, even though there is some quasi-causal relation between quantum events and macro events in actuality, this has no modal force whatever, and hence cannot be considered as causal. Sturgeon calls this inducing: “actual quantum events induce hand movements via composition and causation; and actual hand movements induce quantum events via composition and causation. What fails, at our world, is the link between so inducing and event and causing it” (p. 429) This failure is due to the diversity of our concepts for the macro and the micro: there cannot be modal force — hence no causality — if there is no relation between essences, and this we do not know. There might be, or there might not be: our conceptual apparatus does not justify either position. This simple epistemic point seems sufficient to block the application of the Cause-and-Essence principle. At the same time, it allows us to resist the argument from the generalized no overdetermination premise (iii ) for downward causation: we as yet do not know whether a certain sum of quantum events is sufficient for the instantiation of a given macro property. Macro events, for all we know, may not supervene on micro events. The criticism I want to advance against Sturgeon’s position is quite straightforward. Before passing on to it, however, there are some things to consider about his argument. What Sturgeon has shown is not that quantum events cannot cause
338
AGUST´IN VICENTE
macro events across the divide and vice versa. All he has established, if the rest is conceded, is that we do not know whether there are these “causal diagonals”; as far as we know, there could be. His point must be that there is no way of knowing whether quantum event composition provides the essence of macro events. He cannot deny that they do. From what he says, it only follows that we must suspend our judgment about causal relations between these different realms. A claim for their nonexistence is as unmotivated as is a claim for their existence. Perhaps it is reasonable to be skeptical about a meeting of the concepts of the macro with the concepts of the micro, but even then it does not follow that quantum events are not causing macro events so that there is not a problem about causal exclusion. (Whether there is a problem about explanatory exclusion is quite another matter). Now, the criticism. As said above, Sturgeon starts by distinguishing quantummechanical completeness from broadly physical completeness. Quantum mechanical completeness is true, he says next, but broadly physical completeness is false. Then comes the argument for the skepticism about causal diagonals. However, the following positions are possible: (1) broadly physical completeness is true, and so there is still an argument for physicalism and against mental autonomy; (2) the broadly physical does not have to be causally closed for the mind to be in great trouble: the problems arise as soon as the broadly-physical provides its own candidate for causing behaviors. First: According to what we have said in the introduction, one may think of two ways of justifying the causal closure principle. One is to regard it as a methodological norm. The other is to found it, or a sufficiently powerful version of it, on conservation laws. Now, long before Max Planck started to speak about quanta physicists believed (i) that they were inductively motivated in looking for only physical causes for all physical effects, and (ii) that energy and quantity of movement were conserved. That is, they believed in the causal closure of the physical world as do contemporary physicists. Moreover, this belief about the world of macro physics was perceived to be in conflict with our belief in the causal efficacy of the mental: hence Cartesian interactionism, Malebranche’s occasionalism, etc. Thus, it seemed that if you wanted mental events to cause macro physical events, they had to be physical, otherwise you would be violating what our best theories about the macroscopic world said. These days we think that the best theory of the macro world is Relativity Theory. However, the relativistic description of the macro-world is as causally closed and — apparently — as conceptually remote from quantum mechanics as classical mechanics was. Contra Sturgeon, I would say that there is no room for mental causation in the best description of the macro world available. Second: Suppose for a moment that it is reasonable to doubt that the macrophysical world is causally closed. Then, there is still a problem of exclusion if the macro-effects that have an alleged mental cause have also a macro-physical cause. Now, this seems to be the case of bodily movements. Here is Norman Malcolm stating the problem: “if we bear in mind the comprehensive aspects of the neurophysiological theory — that is, the fact that it provides sufficient causal
THE OVERDETERMINATION ARGUMENT REVISITED
339
conditions for all movements — we shall see that desires and intentions could not be causes of movements” (Malcolm, 1968, p. 136). This is easily converted into an argument for physicalism of a sort, or, at least for naturalism (see fn. 1). The argument, following the lines of the overdetermination argument explained above, establishes that mental properties have to be reduced to neurophysiological properties. And this it does without invoking any principle of causal closure. So Sturgeon has shown that the quantum world and the macro world may lead independent lives. However, he has not shown that the psychological world has this independence from the macro physical world. If mental events cause macro physical events, then mental events have to be macro physical events (neurological or whatever). So we are not very far from where we started, and this in its optimistic interpretation means that the dual explanandum strategy is still open for the antiphysicalist.
2.3. A DISJUNCTIVIST CONCEPTION OF BODILY MOVEMENTS Most philosophers that take the dual explanandum way out of the exclusion problem have it that mental concepts and macro concepts are part of the same worldview, so to speak. They call it common sense. What these philosophers (see, e.g. Rudder Baker, 1995, 2000; Hornsby, 19974 ) defend, then, is that common sense or folk theories cannot be reduced to scientific theories, either in their explanatia or in their explananda and so we are not entitled to make claims for a unique privileged level of causal reality. As a matter of fact, the tendency is to hold that there are two realms, of equal status, of causal reality, but that is, as far as I can see, only supported by the antirealist turn they tend to take. As said above, the presence of a conceptual divide only allows skepticism, neither the claim for causal diagonals nor its denial. Now, it is better to keep macro events and mental events separate. As I have insisted, macro events are also subject to the principle of the causal closure of the physical world: mental events would have a difficult life within that framework. It could be claimed that folk physics is quite another matter, but it is difficult to see how it can be defended that there is, as Sturgeon puts it, a conceptual Grand Canyon between folk physics and scientific physics, especially if we are concerned with movements of objects. Where does the abyss lie here? Therefore, the question I am going to address in this section is whether mental entities, both causes and effects, are essentially related to physical properties (or supervenient on them), or, quite to the contrary, are so remote conceptually that we cannot maintain that both types of events — physical and mental — have common effects. In my discussion I will make use of Jennifer Hornsby’s ideas, for she is someone who has been defending the approach I am now considering very explicitly and for some considerable time.
340
AGUST´IN VICENTE
Hornsby claims to be defending a kind of naturalism, which she calls ‘naive naturalism’. This, however, does not conform to the view taken here about naturalism. As a matter of fact, her arguments are anti-naturalist in the usual sense of ‘naturalism’. This will be a part of her position that I will just skip. Another part that I am not going to consider is her anti-realism privileging explanation over ontology, for we are here concerned with the dual explanandum strategy assuming a minimal realism. From this perspective, from the existence of two types of causal explanations plus a conceptual divide between them only follows skepticism about causal interactions across the divide. Finally, although it is a point she stresses, I will obviate the distinction Hornsby draws between causes and causal-explanations. The idea is that causal relations have events as relata, while causal-explanations can be relations between states. This means that actions do not have causes, but causal-explanations, because their explanantia are states of the entire person, not mental events. This idea introduces a complication in the argument which I think can be set aside for our present purposes. Hornsby’s main point is that there are two different chains of events occurring before and when a human or animal body is moved. First (see the graph in Hornsby, 1997, p. 111) there is the chain starting with the presence of objects of perception; this causes mental events/states of a person; and this, in turn, brings about the desired effects of an action. On the other hand, we have events at sensory surfaces causing brain events/states of a subject of neurophysiological states, and these causing in turn movements of a body. Now according to this view, there are two “moments” at which the chains can touch each other, so to speak. First, it could be argued that actions are movements of a body or some other kind of physical entity; second, it is possible to hold that it is the desired effects of the action which are identical to bodily movements. If either of these views is true, that is, if the chains happen to have some element in common, just the same entities would be involved in one and the other (by the overdetermination argument); hence physicalism would be true. Hornsby’s response to the first possible attack by the physicalist is that actions are neither bodily movements nor any kind of physical goings-on internal to the subject: an action, she says, “is a cause of a bit of the agent’s body’s moving, and [...] is an event of the agent’s trying to do something” (p. 85) Also, “... actions are not to be identified with anything that can be latched onto from an impersonal point of view” (p. 87). Action reports, she says, are essentially transitive — a’s moving of her arm vs. a’s arm moving — and so actions must involve more than just bodily movements: they involve psychological states of a person. So Hornsby’s response to this attack consists basically in individuating actions in a way that is as problematic for the physicalist as are purely mental events or states themselves. But the response is useless if the second attack is successful: what kind of things are the effects of actions? Aren’t they bodily movements? And Hornsby’s answer here is: yes, but of a different sort. There is no unique category of bodily movements: rather, our conception of them is disjunctive. On the one hand,
THE OVERDETERMINATION ARGUMENT REVISITED
341
we have the bodily movements which are effects of neural events; on the other, we have bodily movements caused by actions. Here is how she puts her point: A non-disjunctivist conception of bodily movements, as this is now to be understood, is the basis of much physicalist doctrine. [...] The argument [for physicalism] relies on supposing that there is, as it were, just one bodily movement when there is an action. It holds that we have to identify the causes of actions with the causes of bodily movements discovered by neuro scientists, for we should otherwise have to think that bodily movements are overdetermined. [But] if the disjunctive conception of bodily movements is correct, then a premise of the argument is false. (p. 107) This is just the first step of what she, like any defender of the dual explanandum strategy, needs. But there is more that she needs and does not provide. First, (i) the effects of B-movements cannot at the same time be the effects of A-movements, nor the effects of the latter be the effects of the former, etc. That is, the overdetermination must not be reproducible at any other stage of the chain of causes and effects. Second, (ii) the difference between one kind of bodily movement and the other cannot consist merely in their being effects of different causes. Finally, (iii) it is necessary that bodily movements that are effects of neural events (A-movements) are not sufficient conditions for bodily movements that are effects of actions (Bmovements), and that there is not a relation of essence inclusion between them. Or at least, we need reasons to be skeptical about all this. That is, we need at least that a conceptual Grand Canyon separates the two kinds of bodily movements. These, I claim, are in the end the demands of the dual explanandum way out of the overdetermination argument. 3. A Viable Antiphysicalism? Robert Matthews (1998, p. 893) puts point (i) in this way: “Unless Hornsby is prepared to extend this disjunctive conception throughout the natural world, in what would in effect be a sort of conceptual dualism, there will be explananda for which the two sorts of explanation are in competition. At that point, the usual worries about causal overdetermination arise”. Here Matthews offers, somewhat rhetorically, an escape for the antiphysicalist that some have indeed taken (Rudder Baker, 1995, can be one case). However, these authors tend to believe in causal connections between mental events and behaviors and macro events. This means that the escape would work only if the mental-to-macro connection were not problematic and if the macro were, as it seems to be, conceptually remote from the micro. This, as I have argued, is not the case: the macro world is causally closed in its physical description, and so the “usual worries about causal overdetermination arise” anyway. But maybe the defender of the dual explanandum does not need that much. It is open for her to claim instead that the divergent causal chain stops somewhere
342
AGUST´IN VICENTE
after B-movements are instantiated5 . To get the flavor of what I am saying, we can suppose that the claim is that the mental causal chain interrupts right after Bmovements are instantiated. Or, to be more precise, the mental causal chain runs parallel to the physical causal chain only until B-movements and A-movements are instantiated; at that point, the mental causal chain stops while the physical causal chain goes on. This sounds like epiphenomenalism about B-movements, but it is open for the antiphysicalist to say that B-movements can in turn cause mental events by being perceived6 . Some antiphysicalists would not want to hold exactly this position. Rather, they would claim also that B-movements have further nonphysical effects outside the body of the subject (see the references to the “extended mind” account below). The response to Matthews’ challenge, however, should be, mutatis mutandi, the same as the one I have just offered. Now for condition (ii). B-movements, as will be explained, must be essentially different from A-movements, or at least be conceptually remote from them. What this condition establishes is that this difference in concepts cannot be obtained simply by resort to their different causal antecedents. The difference between Bmovements and A-movements cannot consist merely in that B-movements are caused by mental events while A-movements are caused by neural events. This would beg the question, for what we are trying to establish is precisely whether or not actions or mental events cause something different from A-movements. We cannot start by claiming that they do. What about condition (iii)? A-movements cannot be sufficient for B-movements and B-movements cannot be essential for A-movements. That is, neither the argument from generalized exclusion nor the Cause-and-Essence principle is applicable. Or at least, we need reasons to be skeptical about their applicability. If the first situation obtains, the success of the dual explanandum strategy against the argument from exclusion would be complete. Otherwise, it would be partially successful and only in its negative claims, as the discussion of Sturgeon’s paper has made clear. So something like this is what one may want to hold: Bodily movements are to be approached from two different perspectives. On the one hand, they have to be studied as simple movements of a physical body; on the other, they are movements of a human intentional body. Of course, it must be acknowledged that there is a relation of composition between the two kinds of movement, but it can be said that physical movement composition does not exhaust intentional movements. As happened before with micro and macro movements, it can be said that the movements of a physical body have some features — provided by physical theory — out of which it is difficult to build the defining characteristics of the movements of a human body, for instance (see below) intentionality and expressiveness. Many philosophers will for sure find any account complying with conditions (i–iii) highly implausible. Nonetheless, it is possible to construe a position that is coherent in its commitment to all three conditions. It is a position, moreover, that can be extracted from what some philosophers of cognitive science say, and from
THE OVERDETERMINATION ARGUMENT REVISITED
343
what their vindicated predecessors held. It is not my intention to endorse such a position, only to show that it is not as preposterous as it may seem at first glance. That is, my thesis is that, even if the rest of the ways to evade the argument are denied, the physicalist/antiphysicalist debate is not closed by the overdetermination argument since there is still another way to resist it, and some authors who might endorse such a way. These authors, it must be noted, have not presented their views in response to the overdetermination argument, but as theories of human cognition and agency in general. The application to the overdetermination argument is new and it may seem odd in some ways. More than anything, it may seem odd to develop these views into an antiphysicalist position given that some of the authors that have argued for them want to carry the label of physicalism (for instance, R.A. Wilson, 1995). However, I take it that they are views that, at least, are compatible with antiphysicalism. Authors who are close to the field of robotics, or to what is usually known as “situated cognition”, speak about the necessity to think of the mind as something essentially embodied and embedded. “Classical” cognitive science and philosophy of mind, according to these authors, have been misled by the “computer analogy” and have construed a mistaken view of the human mind. The minimal claim of the defenders of situated cognition is that classical cognitive science has not taken into account the role of the body and of the environment in shaping the human mind. A lot of in principle very demanding processes become rather simple thanks to our having a body by means of which we can act on the environment. Think of playing Tetris solely “in your mind” instead of helping yourself with your hands. Moreover, if you think about it, you will realize not only the importance of our having a body, but also the way we use the environment — in this case, the screen — in order to make our (mental) life simpler. Now some of these authors — Andy Clark (1997) for instance — draw a parallel between their views and those of the phenomenologist Merleau-Ponty. According to this philosopher, and also most probably to the late Husserl, the human body has some distinguishing and essential features that cannot be explained from a physical point of view. The human body is expressive, and it has what Merleau-Ponty calls operative and pre-reflective intentionality (these are concepts related to the idea that the body is capable of understanding the world without having to make use of representations). The human mind is essentially embodied, but it also happens that the human body is essentially mental in that it is the origin, condition of possibility and expression of our mental lives. What has to be kept apart, according to this view, are not minds and bodies, but physical bodies and human bodies. These (sketches of) ideas about the human body can perhaps be made to fulfill the conditions for B-movements explained above. First, (i) it may be argued that there is no reason why B-movements so conceived should interact with the physical world while, at the same time, they may well be able to cause other mental events by being perceived. That is, the movements of the human body are not causes of anything that goes on in the physical chain: all
344
AGUST´IN VICENTE
those events are due to the movements of the physical body. On the other hand, the movements of the human body cause something, since they are perceived (and directly perceived as such) by other subjects. This idea seems to fit Clark and Chalmers’ (1998) theory of the extended mind (see also Wilson’s 1995 “wide computationalism”). According to this idea, the individual mind is not confined to “skin and skull” but rather involves parts of the environment and other minds: the environment contains pieces of information that the mind does not interpret, but directly perceives as such. This information contained in the environment that comes to the mind “as it is” can be said, according to these authors, to be part of the mind. Examples of this phenomenon are the multiplication you have written down on a piece of paper or the screen where you are playing Tetris. These things are accessed by the mind in such a direct way that they cannot be kept apart from it. Now, if these ideas are sensible, then it can be defended that the movements of a human body are part of those parts of the environment which are informationproviders that the mind perceives as such in a direct way. That is, it is possible to “read” the intentions of the movements of a human body just by perceiving them. Second, (ii), (conditional on the success of (iii)) these movements would not differ from the movements of a body considered as physical (A-movements) in terms of their etiology. The difference would be due to their different intrinsic properties. Third, (iii) a conceptual Grand Canyon would protect them from being reduced to a composition of physical events. According to Husserl and especially to Merleau-Ponty, the movements of the human body are peculiar in at least two senses, namely, that they are expressive and that they have a pre-reflective intentionality. It is not very clear what Merleau-Ponty meant by the latter term (see Reuter, 1999), though, as said above, the notion seems close to the idea that the body has some understanding of its environment but does not use full-fledged representations. On the other hand, we have an intuitive grasp of what it is for a bodily movement to be expressive, and the idea that human bodily movements are expressive may be one that we are prima facie prone to accept. Now, one could say that bodies are expressive because they express that they have been moved by some intention, that is, that they are expressive because they express their intentional cause. This, in turn, seems close to saying that the essential difference of human bodily movements is etiological, i.e. that (ii) is not accomplished. However, this is not the idea. For something to express that it has been caused by Q is not the same as it being caused by Q simpliciter. Effects do not usually express their causes. That is, even if ‘F is expressive’ amounted to ‘F expresses its cause’, this could be enough to distinguish human bodily movements from the movements of a physical body. To return to the theories referred to in (i), it can be said that the movements of a human body belong to that part of the environment that cannot be distinguished from the mind in a principled way: they are expressive in that they provide information (maybe about their causal antecedents) that is obtained directly. That is, a way to sum up this point is that human bodily movements are (a) information
THE OVERDETERMINATION ARGUMENT REVISITED
345
gatherers — they have prereflective intentionality — and (b) themselves pieces of information that other minds track in the environment. To close, the defender of this distinction between the human and the physical body could bring in as a way of illustration the distinction between reflexes and intentional movements. Bodily movements which are reflexes may be identical, as far as physical A-movements are concerned, to bodily movements which are not reflexes, that is, that are human bodily movements (B-movements). Therefore, Bmovements do not supervene on A-movements. So the situation here diverges from that of macro to micro relations: it is not just that we have reasons to be skeptical about the truth of supervenience claims. It is that we know them to be false. The same goes for essence relations: we know intentional movements are not essential to physical movements and vice versa7 . 4. Conclusion In this paper I have analyzed a possible response to a powerful argument for mental physicalism, the overdetermination argument. Such a response consists in saying that mental events are causes, but that what they cause are not physical effects. I have tried to show, by criticizing the positions of the two main philosophers who have advanced this response, that this “dual explanandum strategy” is more demanding than it appears to be. However, I have also tried to show that it is not utterly preposterous to follow the strategy up to its last consequences, for there seems to be a philosophical “niche” where it can fit. Acknowledgement Research for this paper has been funded by the BFF2002-03842 project of the Spanish Ministerio de Ciencia y Technología. Notes 1 ‘Naturalism’ and ‘physicalism’ are not synonymous. (Ontological) naturalism would claim that there is nothing
in our world, and in its “natural” counterparts, that is not an entity postulated by the true natural sciences. Physicalism would in this sense be a reductive species of naturalism. This distinction, and the fact that Millikan’s program should be described as naturalist, rather than as physicalist, however, does not change the core of what I say. Indeed, Antony and Levine (1997), for instance, say that the overdetermination argument is the argument for naturalism. 2 In effect, if a mental event is the instantiation of a mental property M by an individual at a time, then a mental event will be identical to a physical event, i.e. the instantiation of a physical property P by an individual at a time, if and only if each of the different terms of the two triads (individual, property, time) are correspondingly identical. In particular, if mental property M is identical to physical property P. If one does not want to take this view on events, and takes a Davidsonian view instead, the argument proceeds in a superficially different but deeply identical way. The premises would be: (i) every physical effect has a complete physical cause that causes it in virtue of being physical; (ii) mental events cause physical changes in virtue of being mental; (iii) no
346
AGUST´IN VICENTE
physical effect is caused by an event in virtue of being mental and in virtue of being physical at the same time; therefore (iv) mental events must be type-identical to physical events. 3 Also Ausonio Marras (1998). 4 See her “Causation in Intuitive Physics and in Commonsense Psychology”. The rest of the papers included in her (1997) point to a divide between the physical in any sense and the mental, but in this paper she seems to commit herself to a causal continuum between mental events and “intuitive physical” events. 5 Nothing that I am saying now and will say in the following paragraphs should be ascribed to Hornsby. I am only supplying possible responses to rebuttals she, or any defender of the dual explanandum strategy, may receive. 6 Precision is needed here. Epiphenomenalism can be taken as claiming that (a) some properties have no causal powers or (b) some properties, or some instantiations of such properties, do not cause anything whatever. What I say about B-movements excludes epiphenomenalism in sense (a) but is compatible with their being epiphenomenal in sense (b). Now a defender of this approach ought to be interested primarily in avoiding the (a) sense, for that kind of epiphenomenalism is a sure road to elimination (see section 1). The (b) sense, though apparently threatening, can be handled more easily. It is not a viable position if beliefs and desires are concerned, for we cannot but hold that beliefs and desires are actual — not merely potential — causes. But we do not have an equivalent strong intuition about the causal efficacy of B-movements. We know that our beliefs and desires cause behaviors, but we cannot say that we know that our behaviors cause further things. So the defender of the strategy has no obligation to prove that B-movements cause, even though the threat of elimination commits her to a defense of their causal potency. 7 There is a third point of divergence between the micro-macro case and this: in a reading of what Sturgeon says about it, there is no inducing relation going upwards here. As Sturgeon explains it, it seems necessary that there is weak supervenience for events of a family to induce, via causation and composition, events belonging to another family. This, so the examples of reflexes show, is not the case for A-movements and B-movements. So the relation seems even weaker than induction.
References Antony, L. and Levine, J. (1997), Reduction with Autonomy, Philosophical Perspectives 11, pp. 83–107. Armstrong, D. (1997), A World of States of Affairs, Cambridge: Cambridge University Press. Baker, L.R. (1995), Explaining Attitudes, Cambridge: Cambridge University Press. Baker, L.R. (2000), Persons and Bodies. A Constitution View, Cambridge: Cambridge University Press. Clark, A. (1997), Being There: Putting Mind, Brain and Body Together Again, Cambridge, MA: MIT Press. Clark, A. and Chalmers, D. (1998), The Extended Mind, Analysis 58, pp. 7–19. Crane, T. and Mellor, H. (1990), There Is No Question of Physicalism, Mind 99, pp. 185–206. Fodor, J. (1974), Special Sciences, or the Disunity of the World as a Working Hypothesis, Synthese 28, pp. 97–115. Hornsby, J. (1997), Simple Mindedness: In Defense of a Naive Naturalism in the Philosophy of Mind, Cambridge, MA: Harvard University Press. Kim, J. (1993), Supervenience and Mind, Cambridge: Cambridge University Press. Kim, J. (1998), Mind in a Physical World, Cambridge, MA.: MIT Press. Levine, J. (2001), Purple Haze: the Puzzle of Consciousness, Oxford: Oxford University Press. Malcolm, N. (1968), The Conceivability of Mechanism, The Philosophical Review 77, pp. 45–72. Marras, A. (1998), Kim’s Principle of Explanatory Exclusion, Australasian Journal of Philosophy 76, pp. 439–451. Matthews, R.J. (1998), Review of Simple Mindedness: In Defense of a Naive Naturalism in the Philosophy of Mind by Jennifer Hornsby, Mind 107, pp. 890–894. Millikan, R. (1993), Explanation in Biopsychology, in White Queen Psychology and Other Essays for Alice, Cambridge: Cambridge University Press. Papineau, D. (1990), Why Supervenience? Analysis 50, pp. 66–71.
THE OVERDETERMINATION ARGUMENT REVISITED
347
Papineau, D. (2000), The Rise of Physicalism, in M. Stone and J. Wolff, eds., The Proper Ambition of Science, London: Routledge. Peacocke, C. (1979), Holistic Explanation, Oxford: Clarendon Press. Sturgeon, S. (1998), Physicalism and Overdetermination, Mind 107, pp. 411–433. Sturgeon, S. (1999), Conceptual Gaps and Odd Possibilities, Mind 108, pp. 377–380. Reuter, M. (1999), Merleau-Ponty’s Notion of Pre-Reflective Intentionality, Synthese 118, pp. 69–88. Wilson, R.A. (1995), Cartesian Psychology and Physical Minds, Cambridge: Cambridge University Press. Yablo, S. (1992), Cause and Essence, Synthese 93, pp. 403–449.
